Open ndellingwood opened 6 months ago
@trilinos/muelu
Hip builds (rocm/5.6.1 on MI210) have the MueLu_CreateOperatorTpetra_MPI_1 failing with diffs between the output and gold files:
MueLu_CreateOperatorTpetra_MPI_1
23:36:56 3/4 Test #689: MueLu_CreateOperatorTpetra_MPI_1 ............................................***Failed Required regular expression not found. Regex=[End Result: TEST PASSED 23:36:56 ] 4.60 sec 23:36:56 ========================================================================================================================== 23:36:56 Xpetra::Parameters 23:36:56 Linear algebra library: Tpetra 23:36:56 Galeri::Xpetra::Parameters<long long> 23:36:56 Matrix type: Laplace2D 23:36:56 Problem size: 10000 (100x100) 23:36:56 Processor subdomains in x direction: 1 23:36:56 Processor subdomains in y direction: 1 23:36:56 Processor subdomains in z direction: -1 23:36:56 ======================================================== 23:36:56 Testing folder "kokkos/" 23:36:56 ========================================================================================================================== 23:36:56 ===================================== Solve 1: Standard ===================================== 23:36:56 --- kokkos/Output/operator_solve_1_np1_tpetra.gold_filtered 2024-05-01 23:35:00.404941557 -0600 23:36:56 +++ kokkos/Output/operator_solve_1_np1_tpetra.out_filtered 2024-05-01 23:35:00.411398308 -0600 23:36:56 @@ -129,6 +129,6 @@ 23:36:56 23:36:56 Smoother (level 2) both : "Ifpack2::Relaxation": {Initialized: true, Computed: true, Type: Symmetric Gauss-Seidel, sweeps: 1, damping factor: 1, Global matrix dimensions: [192, 192], Global nnz: 1682} 23:36:56 23:36:56 -Smoother (level 3) pre : <Direct> solver interface 23:36:56 +Smoother (level 3) pre : KLU2 solver interface 23:36:56 Smoother (level 3) post : no smoother ...
The first diff that pops out is:
23:36:56 -Smoother (level 3) pre : <Direct> solver interface 23:36:56 +Smoother (level 3) pre : KLU2 solver interface
module load python rocm/5.6.1 cmake openmpi/4.1.5 openblas/0.3.23 ninja/1.11.1 export OMPI_CXX=$ROCM_PATH/bin/hipcc export TPETRA_ASSUME_GPU_AWARE_MPI=0
cmake \ -G"Ninja" \ -DCMAKE_INSTALL_PREFIX=$PWD/install \ -DCMAKE_CXX_STANDARD="17" \ -DCMAKE_CXX_COMPILER="which mpicxx" \ -DCMAKE_C_COMPILER="which mpicc" \ -DCMAKE_FORTRAN_COMPILER="which mpifort" \ -DCMAKE_BUILD_TYPE="RELEASE" \ -DBUILD_SHARED_LIBS="OFF" \ \ -DTrilinos_ENABLE_ALL_PACKAGES=OFF \ -DTrilinos_ENABLE_ALL_OPTIONAL_PACKAGES=OFF \ -DTrilinos_ENABLE_EXPLICIT_INSTANTIATION=ON \ -DTrilinos_ASSERT_MISSING_PACKAGES=OFF \ -DTrilinos_ALLOW_NO_PACKAGES=OFF \ -DTrilinos_ENABLE_OpenMP=OFF \ -DTrilinos_ENABLE_TESTS=ON \ \ -DTrilinos_ENABLE_Amesos2=ON \ -DAmesos2_ENABLE_SuperLU=OFF \ -DAmesos2_ENABLE_KLU2=ON \ -DTrilinos_ENABLE_Belos=ON \ -DTrilinos_ENABLE_Ifpack2=ON \ -DTrilinos_ENABLE_Kokkos=ON \ -DKokkos_ARCH_VEGA90A=ON \ -DKokkos_ENABLE_CUDA=OFF \ -DKokkos_ENABLE_HIP=ON \ -DKokkos_ENABLE_OPENMP=OFF \ -DTrilinos_ENABLE_KokkosKernels=ON \ -DTrilinos_ENABLE_MueLu=ON \ -DTrilinos_ENABLE_Tpetra=ON \ -DTpetra_ENABLE_CUDA=OFF \ -DTpetra_INST_HIP=ON \ -DTpetra_INST_SERIAL=OFF \ -DTpetra_INST_OPENMP=OFF \ -DTpetra_INST_DOUBLE=ON \ -DTrilinos_ENABLE_Gtest=ON \ -DTrilinos_ENABLE_Teuchos=ON \ -DTrilinos_ENABLE_Xpetra=ON \ -DTrilinos_ENABLE_Zoltan2=ON \ -DTrilinos_ENABLE_Panzer=ON \ -DTPL_ENABLE_BLAS=ON \ -D BLAS_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \ -D BLAS_LIBRARY_NAMES:STRING="openblas" \ -DTPL_ENABLE_LAPACK=ON \ -D LAPACK_INCLUDE_DIRS:FILEPATH="${OPENBLAS_ROOT}/include" \ -D LAPACK_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \ -D LAPACK_LIBRARY_NAMES:STRING="openblas" \ -DTPL_ENABLE_Netcdf=OFF \ -DTPL_ENABLE_MPI=ON \ -DMPI_USE_COMPILER_WRAPPERS=ON \ -DMPI_EXEC="mpirun" \ -DMPI_EXEC_NUMPROCS_FLAG="-np" \ -DMPI_EXEC_POST_NUMPROCS_FLAGS:STRING="-bind-to;none" \ \ $TRILINOS_DIR
which mpicxx
which mpicc
which mpifort
Automatic mention of the @trilinos/muelu team
Bug Report
@trilinos/muelu
Description
Hip builds (rocm/5.6.1 on MI210) have the
MueLu_CreateOperatorTpetra_MPI_1
failing with diffs between the output and gold files:The first diff that pops out is:
Steps to Reproduce
cmake \ -G"Ninja" \ -DCMAKE_INSTALL_PREFIX=$PWD/install \ -DCMAKE_CXX_STANDARD="17" \ -DCMAKE_CXX_COMPILER="
which mpicxx
" \ -DCMAKE_C_COMPILER="which mpicc
" \ -DCMAKE_FORTRAN_COMPILER="which mpifort
" \ -DCMAKE_BUILD_TYPE="RELEASE" \ -DBUILD_SHARED_LIBS="OFF" \ \ -DTrilinos_ENABLE_ALL_PACKAGES=OFF \ -DTrilinos_ENABLE_ALL_OPTIONAL_PACKAGES=OFF \ -DTrilinos_ENABLE_EXPLICIT_INSTANTIATION=ON \ -DTrilinos_ASSERT_MISSING_PACKAGES=OFF \ -DTrilinos_ALLOW_NO_PACKAGES=OFF \ -DTrilinos_ENABLE_OpenMP=OFF \ -DTrilinos_ENABLE_TESTS=ON \ \ -DTrilinos_ENABLE_Amesos2=ON \ -DAmesos2_ENABLE_SuperLU=OFF \ -DAmesos2_ENABLE_KLU2=ON \ -DTrilinos_ENABLE_Belos=ON \ -DTrilinos_ENABLE_Ifpack2=ON \ -DTrilinos_ENABLE_Kokkos=ON \ -DKokkos_ARCH_VEGA90A=ON \ -DKokkos_ENABLE_CUDA=OFF \ -DKokkos_ENABLE_HIP=ON \ -DKokkos_ENABLE_OPENMP=OFF \ -DTrilinos_ENABLE_KokkosKernels=ON \ -DTrilinos_ENABLE_MueLu=ON \ -DTrilinos_ENABLE_Tpetra=ON \ -DTpetra_ENABLE_CUDA=OFF \ -DTpetra_INST_HIP=ON \ -DTpetra_INST_SERIAL=OFF \ -DTpetra_INST_OPENMP=OFF \ -DTpetra_INST_DOUBLE=ON \ -DTrilinos_ENABLE_Gtest=ON \ -DTrilinos_ENABLE_Teuchos=ON \ -DTrilinos_ENABLE_Xpetra=ON \ -DTrilinos_ENABLE_Zoltan2=ON \ -DTrilinos_ENABLE_Panzer=ON \ -DTPL_ENABLE_BLAS=ON \ -D BLAS_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \ -D BLAS_LIBRARY_NAMES:STRING="openblas" \ -DTPL_ENABLE_LAPACK=ON \ -D LAPACK_INCLUDE_DIRS:FILEPATH="${OPENBLAS_ROOT}/include" \ -D LAPACK_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \ -D LAPACK_LIBRARY_NAMES:STRING="openblas" \ -DTPL_ENABLE_Netcdf=OFF \ -DTPL_ENABLE_MPI=ON \ -DMPI_USE_COMPILER_WRAPPERS=ON \ -DMPI_EXEC="mpirun" \ -DMPI_EXEC_NUMPROCS_FLAG="-np" \ -DMPI_EXEC_POST_NUMPROCS_FLAGS:STRING="-bind-to;none" \ \ $TRILINOS_DIR